How Should a Dashboard Display Sensor Outages?
From the earlier reference collection. Read the current software blog.
Explicit outage markers help readers distinguish missing observations from normal readings and understand the limits of a displayed trend.
Make gaps visible
A missing sensor observation is not a normal value. If a dashboard connects the last reading before an outage to the first reading after it, the smooth line can suggest continuous monitoring that never happened. Leave a visible gap or mark the interval with a distinct outage treatment, and state that readings are unavailable for that period.
Suppose a water sensor reports 18.2 degrees at 09:00, then stops transmitting until 13:00. A line connecting those points may look like a gradual rise or fall. The dashboard should show the four-hour gap and avoid implying what happened in between. If a value was estimated or filled in, distinguish it from the measured readings and explain the method.
Label the status in plain language
Use a status label such as “No data received” or “Sensor offline” when that is what the system can establish. A communication failure does not necessarily prove the sensor itself was broken; it may reflect a network problem, maintenance, or an upstream delay. Choose wording that matches the evidence and avoid diagnosing a cause without support.
Provide the start and end of the known gap, the affected location or device, and the time zone. If the end time is not yet known, say that the outage is ongoing or that recovery has not been confirmed. A brief status summary near the chart can help readers who might not notice a narrow shaded interval.
Keep outages distinct from other data states
A sensor may report a value that fails validation, or a data pipeline may delay a feed even though the device continues collecting. Those conditions differ from a confirmed outage. Use consistent labels for missing, delayed, rejected, and estimated data, and include a legend that explains each state. This prevents a blank area from carrying several incompatible meanings.
In a downloadable table, preserve the timestamp and status field instead of silently dropping the affected rows. A row marked “missing” can be more informative than an absent row, particularly when someone is checking coverage or calculating a summary.
Explain the impact on summaries
Tell readers whether outage periods are excluded from averages, whether the displayed summary is based only on received values, and how much coverage remains. For example, an average based on eight observed hours should not be presented as though it summarizes a complete day. Explicit gaps, careful status language, and a clear coverage note let readers judge whether the remaining data are useful for their question.